Ultrasonic Characterization Method of Carbon Nanotubes Orientation Using Sparse Representation

Carbon nanotubes orientation has a significant impact on the performance of carbon nanotubes/epoxy resin (CNTs/EP) composite materials. The ultrasonic velocity is sensitive to the degree of carbon nanotubes orientation. However, the time-of-flight (TOF) is difficulty to accurately measured due to ultrasonic echo aliasing and waveform distortion occurred in the thin CNTs/EP composite material. Therefore, the ultrasonic characterization method of carbon nanotubes orientation based on the sparse representation is studied in this paper. Two groups of CNTs/EP specimens are taken as research objects. Firstly, the dictionary is constructed by Gabor function, and the matching degree between dictionary and ultrasonic echo signals is researched. Then, the matching pursuit (MP) algorithm is used to solve the problem of dictionary redundancy and obtain the best matching atom; the signal reconstruction is completed and ultrasonic TOF of carbon nanotubes is extracted more accurately. Finally, a carbon nanotubes orientation prediction model based on the sparse representation is established. The result shows that the Gabor dictionary has good similarity with ultrasonic echo signals and the root mean square relative error is 0.0147. The degree of carbon nanotubes orientation and ultrasonic velocity have positive correlation. The carbon nanotubes orientation prediction model relative error is 5.58%. This paper has provided an effective method to characterize the degree of carbon nanotubes orientation.